The Restoration of Textured Images Using Fractional-Order Regularization

Algorithm 1

Fractional-order regularization.

Input: observed image 𝐲, PSF 𝐡, penalty parameters 𝜆, exponent 𝛼 and fractional-order 𝑣

Input: 𝛽₀, 𝛽_Max, 𝜃₀, 𝜃_Max

Initialize 𝐱=𝐲, 𝜃=𝜃₀, 𝛽=𝛽₀

while 𝜃<𝜃_Max, do

 solve for 𝑢 using (12)

while 𝛽<𝛽_Max do

 Given 𝐱, solve for 𝐰 according to our discussion

 Given 𝐰, solve for 𝐱 using (10)

  𝛽=2𝛽

end while

  𝜃=2𝜃

end while

Output: Estimated image 𝐱